TL;DR
This study directly compares two common lab methods for inducing forgetting: directed forgetting (intentional) and recognition-induced forgetting (incidental). They found that recognition-induced forgetting leads to a larger forgetting effect and may be a more robust method for "use-inspired" forgetting research, suggesting practical differences in how we forget.
What Problem Does This Paper Address?
For decades, researchers have sought reliable ways to induce forgetting in the lab, aiming to understand how we control what we remember and forget. Two prominent paradigms are directed forgetting, where participants are explicitly told to remember or forget specific items, and recognition-induced forgetting, where remembering one item incidentally causes us to forget related items. While both are established, it's unclear which is more effective or if they rely on the same underlying mechanisms. This paper aims to directly compare these two methods to see which produces a stronger forgetting effect and whether they are correlated, offering insights into how forgetting might function in real-world scenarios.
Methods
The researchers conducted two experiments using a within-subjects design, meaning each participant experienced both directed forgetting and recognition-induced forgetting tasks.
- Directed Forgetting: Participants were shown images and given cues to either "remember" or "forget" them. Later, their memory for these items was tested. This paradigm relies on participants' intentional control.
- Recognition-Induced Forgetting: This was assessed by comparing memory for "Baseline" items (remembered and not related to other items) with "Related" items (remembered but related to other items that were also remembered). The idea is that remembering a "target" item (which was cued for remembering) leads to forgetting of related items. This is considered an incidental process.
Experiment 2 introduced modifications to address potential issues in Experiment 1:
- Additional Response Options: Participants had more choices during the test phase to clarify their understanding of the "forget" cue, aiming to reduce confusion.
- Post-Experiment Survey: Participants were asked if they expected "forget" items to be tested. This allowed the researchers to exclude participants who might have disregarded the "forget" cue due to distrust, providing a "purer" measure of directed forgetting.
Memory performance was primarily measured using the Area Under the Curve (AUC) from signal detection theory, which accounts for both accuracy and confidence. They also reported hit rates and d' for comparison.
Datasets
The study used a custom stimulus set consisting of 15 object categories, with 18 exemplars per category. These stimuli were made available on the Open Science Framework (OSF). No new benchmarks were introduced; rather, the study utilized existing paradigms and evaluated their comparative effectiveness.
Key Insights and Ideas
- Forgetting is Not Monolithic: The study highlights that different laboratory methods can induce forgetting, and these methods may not be interchangeable.
- Incidental Forgetting Can Be Stronger: Recognition-induced forgetting, which happens unintentionally, appears to be a more potent forgetting mechanism than directed forgetting, which relies on intentional control.
- Demand Characteristics Matter: Experiment 2's findings suggest that participants' expectations about whether "forget" cues will actually lead to items being tested can significantly influence directed forgetting effects. This implies that directed forgetting paradigms might be susceptible to participants trying to "game" the experiment.
- Potential for Shared Mechanisms (with caveats): While Experiment 1 showed no correlation between the two forgetting types, Experiment 2, with improved methodology, found a reliable correlation. This suggests that while distinct, they might tap into some shared underlying memory processes, though further research is needed to confirm this.
Key Results
- Both paradigms induce forgetting: In both experiments, participants showed reliable forgetting for "forget" items compared to "baseline" items (directed forgetting) and for "related" items compared to "baseline" items (recognition-induced forgetting).
- Recognition-induced forgetting is larger: In Experiment 1, there was no significant difference in the magnitude of forgetting between the two methods. However, in Experiment 2, after controlling for potential confounds, recognition-induced forgetting was significantly larger than directed forgetting across multiple measures (AUC, hit rate, d').
- Correlation findings: Experiment 1 found no reliable correlation between directed and recognition-induced forgetting. Experiment 2, however, found a significant positive correlation, suggesting that individuals who exhibit stronger recognition-induced forgetting also tend to show stronger directed forgetting.
- Participant distrust: In Experiment 2, nearly half of the participants reported believing that "forget" items would be tested, suggesting a potential issue with the directed forgetting paradigm's reliance on participant compliance.
Limitations and Future Work
- Demand Characteristics in Directed Forgetting: The authors acknowledge that directed forgetting might be susceptible to demand characteristics and response confusion, as evidenced by the results of Experiment 2. This makes it harder to isolate the "pure" intentional forgetting effect.
- Shared Mechanism Nuance: While Experiment 2 found a correlation, the authors suggest that further work is needed to confirm a shared underlying mechanism, perhaps by comparing two directed forgetting tasks or two recognition-induced forgetting tasks within the same participants.
- Robustness of Recognition-Induced Forgetting: The paper argues that recognition-induced forgetting might be a more robust method for use-inspired forgetting research due to its larger effect size and less susceptibility to participant expectations.
- Real-World Generalizability: The authors suggest that understanding these differences is crucial for real-world applications, such as eyewitness testimony or clinical interventions for conditions like PTSD, where controlling memory is important.
Summary updated